JP5732773B2 - Image identification device, image identification method, and program - Google Patents

Description

  The present invention relates to an image identification device, an image identification method, and a program.

  Conventionally, a focusing technique based on a focus evaluation value obtained by detecting a human face has been known (see, for example, Patent Documents 1 and 2).

JP 2006-201282 A JP-A-8-63597

  However, in Patent Documents 1 and 2, since the face is detected from the entire image area of the live view image, there is a problem in that the amount of calculation increases and it takes time to identify the face.

  Accordingly, an object of the present invention is to provide an image identification device, an image identification method, and a program capable of increasing the identification speed of a specific image region.

In order to solve the above problems, an image identification apparatus according to a first aspect of the present invention includes a first recording unit that records identification information for identifying a specific image region, and a second recording unit that records an image. Generating means for generating a plurality of first areas and second areas from the image recorded in the second recording means; the image recorded in the second recording means; and the first recording means Calculating means for calculating the similarity to the identification information recorded in the first information, and calculating the similarity between each of the first areas generated by the generating means and the identification information. Based on the similarity calculated by the first calculation control means and the calculation control means, the similarity between the identification information and the second area generated by the generation means is calculated. The area to be calculated by means Specifying means for determining, a second calculation control means for calculating the similarity between the area specified by the specifying means and the identification information, the first calculation control means, and the second calculation control means. Based on the similarity calculated by at least one of the calculation control means, the identification means for identifying the specific image region in the image, and the similarity calculated by the first calculation control means, Determination means for determining whether or not the calculation means calculates the similarity with the identification information for each of the second areas based on the similarity of the plurality of first areas around the second area And the designation means designates an area in which the calculation means calculates the similarity to the identification information among each of the second areas based on a determination result by the determination means. Special It is set to.

According to a second aspect of the present invention, in the image identification device according to the first aspect,
Based on the similarity calculated by the first calculation control means, the information processing apparatus further comprises a determination means for determining whether the calculation means calculates the similarity with the identification information for each of the second regions. The designation means designates an area in which the calculation means calculates the similarity with the identification information among each of the second areas based on the determination result by the determination means.

According to a third aspect of the present invention, in the image identification device according to the second aspect,
And further comprising a setting means for setting a threshold value as a determination criterion by the determination means as to whether or not the similarity with the identification information for each of the second areas is calculated by the calculation means, Further, when it is determined that the similarity calculated by the first calculation control means is equal to or greater than the threshold set by the setting means, it is determined that the similarity with the identification information is calculated by the calculation means. On the other hand, when it is determined that the degree of similarity calculated by the first calculation control unit is less than the threshold set by the setting unit, the degree of similarity with the identification information is calculated by the calculation unit. It is characterized by not judging.

The invention according to claim 4 is the image identification device according to any one of claims 1 to 3,
The image recording apparatus further includes image reduction means for sequentially reducing the images recorded in the second recording means at a predetermined ratio to sequentially generate reduced images, and the generation means includes the reduced images sequentially generated by the image reduction means. The first region and the second region are generated from each of the first and second regions, respectively.

The invention according to claim 5 is the image identification device according to any one of claims 1 to 4,
The generating unit generates the first area and the second area so that a reference block including predetermined unit pixels of the first area and the second area is divided into direct sums.

The invention according to claim 6 is the image identification device according to any one of claims 1 to 5,
The image processing apparatus further includes display means for displaying an image recorded in the second recording means.

The invention according to claim 7 is the image identification device according to claim 6,
The image processing apparatus further includes display control means for performing predetermined processing on the image corresponding to the specific image area identified by the identification means in the image and displaying the image on the display means.

The invention according to claim 8 is the image identification device according to any one of claims 1 to 7,
The image recording apparatus further includes an image capturing unit that captures an image, and the second recording unit records the image captured by the image capturing unit.

An image identification method according to a seventh aspect of the invention includes a first recording unit that records identification information for identifying a specific image area, a second recording unit that records an image, and the second recording unit. An image identification method using an image identification apparatus comprising: a calculation unit that calculates a similarity between an image recorded in the unit and the identification information recorded in the first recording unit; A generation step of generating a plurality of first areas and a plurality of second areas from the image recorded in the second recording means, and the similarity between each of the generated first areas and the identification information is calculated. Based on the first calculation control step calculated by the means and the similarity calculated in the first calculation control step, the similarity with the identification information in each of the second regions is calculated. Specify the area to be calculated by means Of the first step and the second region, a specifying step, a second calculation control step for calculating the similarity between the region specified in the specifying step and the identification information by the calculating means, A degree of similarity calculated by the first calculation control step and the step of identifying the specific image area in the image based on the similarity degree related to at least one of the second area and the surrounding area of the second area A determination step for determining whether or not the calculation means calculates a similarity to the identification information for each of the second areas based on a plurality of similarities of the first areas, and including the designation step, based on a determination result by the determination step, of each of the second region, especially to specify the region for calculating the degree of similarity between the identifying information by said calculation means It is set to.

According to an eighth aspect of the present invention, there is provided a first recording means for recording identification information for identifying a specific image area, a second recording means for recording an image, and a second recording means. A computer of an image identification device, comprising: a calculation unit that calculates a similarity between the recorded image and the identification information recorded in the first recording unit; Generating means for generating a plurality of first areas and a plurality of second areas from an image, and a first means for calculating the similarity between each of the first areas generated by the generating means and the identification information. Calculation control means, and based on the similarity calculated by the first calculation control means, the calculation means calculates the similarity to the identification information in each of the second regions generated by the generation means. Calculated by Designation means for designating an area; second calculation control means for calculating the similarity between the area designated by the designation means and the identification information; the first calculation control means; and the second calculation control means. Identification means for identifying the specific image region in the image based on the similarity calculated by at least one of the calculation control means, the similarity calculated by the first calculation control means, Determination means for determining whether or not the calculation means calculates the similarity with the identification information for each of the second areas based on the similarity of the plurality of the first areas around the second area; to function as the designation means, on the basis of the determination result by the determination means, of each of the second region, designates a region to be calculated by the calculation means the similarity between the identification information It is characterized by a door.

  According to the present invention, it is possible to increase the identification speed of a specific image region.

It is a block diagram which shows schematic structure of the image display apparatus of one Embodiment to which this invention is applied. 3 is a flowchart illustrating an example of an operation related to image identification processing by the image display apparatus of FIG. 1. 3 is a flowchart illustrating an example of an operation related to a face detection process in the image identification process of FIG. 2. It is a figure which shows typically an example of the image which concerns on the image identification process of FIG. It is a figure which shows typically an example of the identification object of the face which concerns on the image identification process of FIG. It is a figure which shows typically an example of the 1st area | region as an identification object of FIG. It is a figure which shows typically an example of a structure of the sub discrimination device of the evaluation value calculation part which concerns on the image identification process of FIG. It is a figure which shows typically an example of the evaluation value of the 2nd area | region which concerns on the image identification process of FIG. It is a figure which shows typically an example of the face area | region which concerns on the image identification process of FIG. It is a block diagram which shows schematic structure of the imaging device of one Embodiment to which this invention is applied.

Hereinafter, specific embodiments of the present invention will be described with reference to the drawings. However, the scope of the invention is not limited to the illustrated examples.
FIG. 1 is a block diagram showing a schematic configuration of an image display apparatus 100 according to an embodiment to which the present invention is applied.

The image display apparatus 100 according to the present embodiment generates two types of image areas from the image, and identifies each of the other image areas based on the similarity between each of the image areas and the identification information. Specify an image area for calculating the similarity to the information for use, and specify in the image P based on the calculated similarity for at least one of the image area and the other image area An image area (eg, a face area) is identified.
Specifically, the image display device 100 is, for example, a digital photo frame installed on a table, and as shown in FIG. 1, a display unit 1, an image processing unit 2, a face detection unit 3, and a memory 4, an identification information recording unit 5, an image recording unit 6, an operation input unit 7, and a central control unit 8.

The display unit 1 includes a display panel 1a and a display control unit 1b.
The display panel 1a constitutes display means for displaying an image recorded in the image recording unit 6 (for example, the image P or the like; see FIG. 9B). Examples of the display panel 1a include a liquid crystal display panel and an organic EL display panel. However, the display panel 1a is an example and is not limited thereto.
The display control unit 1b displays a predetermined image on the display screen of the display panel 1a based on the image data of a predetermined size read from the image recording unit 6 and decoded by the image processing unit 2.

In addition, the display control unit 1b performs predetermined processing on the face area as the specific image area identified by the face identification unit 3i (described later) in the image, and causes the display panel 1a to display the face area. Specifically, for example, the display control unit 1b enlarges and displays an image region corresponding to the face region identified by the face identifying unit 3i in the image recorded in the image recording unit 6 at a predetermined magnification. It is displayed on the display screen of the panel 1a (not shown). Further, for example, the display control unit 1b performs demosaic processing on the image area corresponding to the face area identified by the face identification unit 3i in the image recorded in the image recording unit 6 to display on the display panel 1a. It is displayed on the screen (not shown).
Further, when a plurality of face areas are identified by the face identifying section 3i, the display control section 1b performs predetermined processing on the image areas corresponding to the sequentially identified face areas, and then these images are displayed. The areas are switched at a predetermined display time and displayed on the display screen of the display panel 1a.
In addition, what was illustrated as a predetermined | prescribed process is an example, and is not restricted to these.
In this way, the display control unit 1b configures display control means that performs predetermined processing on the image area corresponding to the specific image area identified by the identification means in the image and displays the image area on the display panel 1a. Yes.

The image processing unit 2 is a predetermined encoding method (for example, JPEG format, motion JPEG format, MPEG format, etc.) corresponding to the image data or moving image data of the image P related to the display target read from the image recording unit 6. Are decoded and output to the display control unit 1b. The image processing unit 2 reduces the image data read from the image recording unit 6 to a predetermined size (for example, VGA or QVGA size) based on the display resolution of the display panel 1a, for example, and the display control unit. Output to 1b.
Further, the image processing unit 2 generates reduced image data R (see FIGS. 4A to 4F) from the luminance image data L of the image P related to the display object in the image identification process. Part 2a.

  Based on the luminance image data L of the image P relating to the display target read from the image recording unit (second recording unit) 6, the reduced image generating unit 2 a is both horizontal (x axis) and vertical (y axis). Reduced image data R reduced at a predetermined ratio is generated. Specifically, the reduced image generation unit 2a sequentially sets the horizontal and vertical pixels of the image data in a predetermined ratio (for example, 0.9 times) sequentially based on the luminance image data L of the image P related to the display target. Reduced image data R whose resolution is reduced stepwise by reduction is sequentially generated (see FIGS. 4A to 4F).

Note that the number of reduced image generations (number of reductions) can be arbitrarily changed as appropriate in consideration of the size of input luminance image data, the identification accuracy and identification speed of a specific image region (face region), and the like.
As described above, the reduced image generation unit 2a constitutes an image reduction unit that sequentially reduces the images recorded in the image recording unit 6 at a predetermined ratio to sequentially generate reduced images.

  The face detection unit 3 includes an image region generation unit 3a, a temporary storage memory 3b, an evaluation value calculation unit 3c, a first calculation control unit 3d, a determination unit 3e, an evaluation threshold setting unit 3f, and an image region designation 3g, the 2nd calculation control part 3h, and the face identification part 3i are comprised.

The image area generation unit 3a generates a first area A1 and a second area A2 from the image P related to the display target.
That is, the image area generation unit 3a uses the first area A1 and the second area to be face identification target areas by the evaluation value calculation unit 3c from the luminance image data L of the image P related to the display target generated by the image processing unit 2. A plurality of areas A2 are generated (see FIG. 5). Specifically, the image area generation unit 3a has a predetermined size (for example, so that a reference block composed of predetermined unit pixels (for example, one pixel) of each of the first area A1 and the second area A2 is divided into direct sums. , 24 × 24 pixels) first area A1 and second area A2.

For example, when the pixel in the upper left corner (predetermined position) of the first area A1 and the second area A2 is used as a reference block, the image area generation unit 3a is configured as shown in FIG. 6 (a), FIG. The reference blocks of the first area A1 and the second area A2 are set to be arranged in a checkered pattern. Here, in the present embodiment, the reference block of the first area A1 corresponds to a checkered black portion, and the reference block of the second area A2 corresponds to a checkered white portion (FIG. 8). Etc.). Then, the image area generation unit 3a has a predetermined size (for example, 24 × 24 pixels) sub-window A (see FIG. 5) with reference to each reference block related to the first area A1 in the luminance image data L and the reduced image data R. The first region A1 is sequentially generated by setting a plurality of to predetermined positions. Then, after the generation of the first area A1, the image area generation unit 3a, like the first area A1, has a predetermined size with reference to each reference block related to the second area A2 in the luminance image data L and the reduced image data R. By setting a plurality of sub-windows A (for example, 24 × 24 pixels) at predetermined positions, the second area A2 is sequentially generated.
Specifically, for the first area A1, the image area generation unit 3a uses, for example, a pixel at coordinates (0, 0) as a reference block, and (0, 0), (23, 0), (0 , 23), (23, 23) are the first area A1 (see FIG. 6A) having the coordinates of the four corners, the pixel at the coordinates (2, 0) is the reference block, and (2, 0), The first area A1 (see FIG. 6B) having (25, 0), (2, 23), (25, 23) as the coordinates of the four corners, and the pixel at the coordinates (4, 0) as the reference block, In addition, a first area A1 (see FIG. 6C) having four corner coordinates as (4, 0), (27, 0), (4, 23), and (27, 23) is sequentially generated. In addition, when the image area generation unit 3a generates the first area A1 to the right end in the luminance image data L and the reduced image data R, the setting position of the sub window A is shifted by one pixel in the y-axis direction, A first area A1 is generated from left to right.
Similarly, for the second area A2, the image area generation unit 3a uses, for example, a pixel at coordinates (1, 0) as a reference block, and (1, 0), (24, 0), (1, 23 ), (24, 23) as the second block A2 (not shown) having the coordinates of the four corners, and the pixel at the coordinates (3, 0) as the reference block, and (3, 0), (26, 0), The second area A2 (not shown) having the coordinates of (3, 23), (26, 23) and the pixel at the coordinates (5, 0) as a reference block, and (5, 0), (28 , 0), (5, 23), (28, 23) are sequentially generated as a second region A2 (not shown) having four corner coordinates. In addition, when the image region generation unit 3a generates the second region A2 to the right end in the luminance image data L and the reduced image data R, the setting position of the sub window A is shifted by one pixel in the y-axis direction, A second area A2 is generated from left to right.
Although illustration is omitted, the image area generation unit 3a determines that the coordinates of the pixels in the lower right corner of the first area A1 and the second area A2 are the coordinates of the pixels in the lower right corner of the luminance image data L and the reduced image data R. Or the generation of the first area A1 and the second area A2 is terminated. In addition, the generation direction (generation order) of the first area A1 and the second area A2 is an example and is not limited to this, and can be arbitrarily changed as appropriate.

The image area generator 3a generates a plurality of first areas A1 and a plurality of second areas A2 from each of the reduced image data R of the luminance image sequentially generated by the reduced image generator 2a. That is, every time the reduced image data R is sequentially generated by the reduced image generator 2a, the image area generator 3a generates a plurality of first areas A1 and second areas A2 from the reduced image data R, respectively.
In this way, the image area generation unit 3a constitutes a generation unit that generates the first area A1 and the second area A2 from the image recorded in the image recording unit 6, respectively.

The temporary storage memory 3b temporarily stores the image data of the first area A1 and the second area A2 generated by the image area generation unit 3a. That is, when the first area A1 and the second area A2 are sequentially generated by the image area generator 3a, the image data of the first area A1 and the second area A2 are sequentially stored in the temporary storage memory 3b.
Further, the temporary storage memory 3b uses the similarity evaluation value related to the similarity between the identification target area such as the first area A1 and the second area A2 calculated by the evaluation value calculation unit 3c and the identification information as the identification target area. Temporarily store it in association.

The evaluation value calculation unit 3 c calculates a similarity evaluation value related to the similarity between the display target image and the identification information recorded in the identification information recording unit 5.
That is, the evaluation value calculation unit 3c acquires the image data of the first region A1 and the second region A2 generated by the image region generation unit 3a from the temporary storage memory 3b, respectively. A similarity evaluation value related to the similarity between the area A2 and the identification information is calculated. Further, the evaluation value calculation unit 3c calculates a similarity evaluation value by using, for example, adaboost output calculation for the image data of the first region A1 and the second region A2 that are input. Specifically, the evaluation value calculation unit 3c includes a plurality of sub classifiers for determining whether the identification target area such as the first area A1 or the second area A2 is a face. A plurality of sub classifiers are defined in a plurality of (for example, 20) stages. For example, the sub-classifier unit has a predetermined number of sub-classifiers defined in each stage, such as two in the first stage, five in the second stage, ten in the third stage, and twenty in the fourth stage. Yes. Note that the sub-classifier defined for a lower (smaller numerical value) stage may be set to have a higher reliability in determining whether or not it is a face.
Then, the evaluation value calculation unit 3c inputs the image data of the identification target area to the sub-classifiers of each stage in the order of stages. The evaluation value calculation unit 3c inputs image data ("T" in FIG. 7) determined to be a face by all the sub-classifiers of one stage to the sub-classifier of the next stage, The subsequent identification of the image data determined to be not (“F” in FIG. 7) is stopped. The evaluation value calculation unit 3c calculates a similarity evaluation value for each identification target area such as the first area A1 and the second area A2 according to the identification result of the sub-classifier. Specifically, the evaluation value calculation unit 3c determines, for each identification target region, the number of passages of a plurality of stages that define a plurality of sub-classifiers, that is, faces at all sub-classifiers defined in each stage. If it is determined that the number of sub-discriminators specified in all the stages that have passed through the stage is passed, the value is calculated as the similarity evaluation value. To do. On the other hand, when the number of passes through the stage is “0” (when the stage cannot be passed through the first stage and is not delivered to the second stage), the evaluation value calculation unit 3c, for example, has a predetermined minimum value such as a negative value Is calculated as a similarity evaluation value.

For example, as shown in FIG. 7, the evaluation value calculation unit 3c inputs the image data of the identification target area to the two sub-classifiers in the first stage in a predetermined order, and both of these two sub-classifiers The image data of the identification target area determined to be a face is input to the second stage sub-classifier, which is the next stage. On the other hand, if it is determined that the similarity calculated by the evaluation value calculation unit 3c up to the second sub classifier of the first stage is less than a preset threshold, the evaluation value calculation unit 3c The subsequent identification is stopped, and a predetermined negative value is calculated as the similarity evaluation value. Specifically, for example, the first stage sub-identifier determines an image area with little luminance change, such as a blue sky or a plain background part. In the first stage, a blue sky or a plain background part is used. If it is determined that the image area has little change in luminance, the identification of whether or not the face is a subsequent face is stopped.
Further, the evaluation value calculation unit 3c applies the image data of the identification target area determined to be a face in the first stage to the five sub-classifiers in the second stage in a predetermined order as in the first stage. The input image data for which the similarity calculated from the first sub-classifier to the fifth sub-classifier is determined to be greater than or equal to a preset threshold value is input to the third stage, which is the next stage. Input to the classifier. On the other hand, if the similarity calculated by the evaluation value calculation unit 3c from the first sub-classifier of the second stage to the fifth sub-classifier is determined to be less than a preset threshold value, the evaluation value is calculated. The unit 3c stops the subsequent identification of the image data, and the total number of sub classifiers up to the previous stage (first stage) that has passed (if only the first stage passes, “2 ]) As a similarity evaluation value.
The evaluation value calculation unit 3c calculates the similarity in the same manner as described above for the third stage and thereafter. That is, an image region having a higher possibility of being determined to be a face is input to the sub-classifier at the earlier stage.

Note that the method of calculating the similarity by the evaluation value calculation unit 3c is an example, and is not limited to this, and can be arbitrarily changed as appropriate.
As described above, the evaluation value calculation unit 3 c constitutes a calculation unit that calculates the similarity between the image recorded in the image recording unit 6 and the identification information recorded in the identification information recording unit 5. .

The first calculation control unit 3d performs control to cause the evaluation value calculation unit 3c to calculate a similarity evaluation value related to the similarity between each of the first areas A1 and the identification information recorded in the identification information recording unit 5. .
That is, the first calculation control unit 3d controls the evaluation value calculation unit 3c every time the first region A1 is generated by the image region generation unit 3a, and generates the generated first regions A1 and the identification information. The similarity evaluation value related to the similarity is calculated.
Here, the first calculation control unit 3d includes a first calculation control unit that causes the evaluation value calculation unit 3c to calculate the similarity between each of the first regions A1 generated by the image region generation unit 3a and the identification information. It is composed.

The determination unit 3e determines whether or not the evaluation value calculation unit 3c calculates the similarity evaluation value related to the similarity with the identification information for each of the second regions A2.
That is, the determination unit 3e relates to the similarity with the identification information for each of the second regions A2 based on the similarity evaluation value calculated by the evaluation value calculation unit 3c under the control of the first calculation control unit 3d. It is determined whether or not the similarity evaluation value is calculated by the evaluation value calculation unit 3c. Specifically, for each of the second areas A2 serving as identification target areas, the determination unit 3e is similar to at least one first area A1 including a reference block adjacent to the reference block of the second area A2. Based on the evaluation value, it is determined whether or not the evaluation value calculation unit 3c calculates the similarity evaluation value of the identification target region (second region A2). More specifically, the determination unit 3e is similar to each first area A1 including a reference block adjacent to the reference block of the second area A2 serving as the identification target area among the plurality of second areas A2,. A value obtained by adding the evaluation values is used as a determination evaluation value for the second region A2, and it is determined whether or not the evaluation threshold value is greater than or equal to the evaluation threshold set by the evaluation threshold setting unit 3f. When the determination unit 3e determines that the evaluation value for determination related to the first region A1 is equal to or greater than the evaluation threshold set by the evaluation threshold setting unit 3f, the evaluation value calculation unit 3c calculates the similar evaluation value. to decide. On the other hand, when determining that the evaluation value for determination related to the first region A1 is less than the evaluation threshold, the determination unit 3e determines that the evaluation value calculation unit 3c does not calculate a similar evaluation value.
For example, as shown in FIG. 8, when four reference blocks of the first area A1 are adjacent to the reference block of the second area A2 that is the identification target area, the determination unit 3e A value obtained by adding all the similar evaluation values (for example, “2”, “2”, “7”, “7”, etc.) of the first region A1 related to the second region A2 is equal to or higher than the evaluation threshold value. It is determined whether or not. When the determination unit 3e determines that the evaluation value for determination is equal to or greater than the evaluation threshold value, the determination unit 3e determines that the evaluation value calculation unit 3c calculates a similar evaluation value for the second region A2 that is the identification target region. When it is determined that the evaluation value for determination is less than the evaluation threshold value, it is determined that the evaluation value calculation unit 3c does not calculate the similar evaluation value for the second region A2 that is the identification target region.

An example of the evaluation value for determination of the second area A2 as the identification target area is a value obtained by adding the similarity evaluation values of the first areas A1 including the reference block adjacent to the reference block of the second area A2. However, this is merely an example, and the present invention is not limited to this. For example, the average value, maximum value, minimum value, median value, etc., of the similar evaluation values of the plurality of first regions A1,.
In this manner, the determination unit 3e determines whether or not the evaluation value calculation unit 3c calculates the similarity with the identification information for each of the second regions A2 based on the similarity calculated by the first calculation control unit 3d. It constitutes a judging means for judging whether or not.

The evaluation threshold value setting unit 3f sets an evaluation threshold value related to the determination criterion by the determination unit 3e.
That is, the evaluation threshold value setting unit 3f is an evaluation that is a determination criterion by the determination unit 3e as to whether or not the evaluation value calculation unit 3c is to calculate a similarity evaluation value related to the similarity between each of the second areas A2 and the identification information. Set the threshold. Specifically, the evaluation threshold value setting unit 3f determines a predetermined value input based on a predetermined operation of the operation input unit 7 by the user or a predetermined value set in advance as a default by the determination unit 3e. Set as a reference evaluation threshold.
Here, the evaluation threshold value setting unit 3f sets a threshold value that serves as a determination criterion by the determination unit 3e as to whether or not the evaluation value calculation unit 3c is to calculate the similarity to the identification information for each of the second areas A2. It constitutes setting means.

The image area designation unit 3g designates an image area in which the evaluation value calculation unit 3c calculates the similarity evaluation value related to the similarity to the identification information among the plurality of second areas A2,.
That is, the image area designating unit 3g evaluates the similarity evaluation value of each of the second areas A2 based on the similarity evaluation value calculated by the evaluation value calculation unit 3c under the control of the first calculation control unit 3d. An image region to be calculated by the value calculation unit 3c is designated. Specifically, as a result of the determination by the determination unit 3e, the image region specifying unit 3g uses the evaluation value calculation unit 3c to determine the similarity evaluation value related to the similarity to the identification information among the plurality of second regions A2,. The second area A2 that is determined to be calculated is designated as an image area for calculating the similarity evaluation value. For example, the image area designating unit 3g uses the similarity evaluation values (for example, “2”, “2”, “7” of the first area A1 related to the reference block adjacent to the reference block of the second area A2 serving as the identification target area. ”,“ 7 ”, etc.) is designated as the image area for calculating the similarity evaluation value, for which the value (evaluation value for determination of the second area A2) determined to be equal to or greater than the evaluation threshold value. .
Here, the image area designating unit 3g is similar to the identification information among the second areas A2 generated by the image area generating unit 3a based on the similarity calculated by the first calculation control unit 3d. A designating means for designating an image area in which the degree is calculated by the evaluation value calculation unit 3c is configured.

The second calculation control unit 3h performs control to cause the evaluation value calculation unit 3c to calculate a similarity evaluation value related to the similarity between the image region specified by the image region specifying unit 3g and the identification information.
That is, the second calculation control unit 3h includes, among the second regions A2 generated by the image region generation unit 3a, the second region A2 specified as the image region by the image region specification unit 3g and the identification information. The similarity evaluation value related to the similarity is calculated by the evaluation value calculation unit 3c.
Here, the second calculation control unit 3h constitutes a second calculation control unit that causes the evaluation value calculation unit 3c to calculate the similarity between the image region specified by the image region specifying unit 3g and the identification information. .

The face identifying unit 3i identifies a face area in an image as a display target.
That is, the face identification unit 3i includes the similarity evaluation value of each first region A1 calculated by the first calculation control unit 3d and the similarity evaluation value of each second region A2 calculated by the second calculation control unit 3h. Based on at least one, a face area (specific image area) is identified in a predetermined image. Specifically, the face identification unit 3i evaluates the similarity between the first region A1 and the second region A2 generated from the luminance image of the image to be displayed and the reduced image generated by the reduced image generation unit 2a. Based on the value, the face area is identified (see FIG. 9A). FIG. 9A schematically shows a state in which each temporary face frame Wa is superimposed on each face area identified in the luminance image and each reduced image. Wa may be displayed on the display image displayed on the display panel 1a or may not be displayed.
Then, the face identifying unit 3i integrates at least one face area identified from the luminance image or the reduced image according to a predetermined arithmetic expression, and specifies an image area corresponding to the one face area in the display target image. (See FIG. 9B).

Although the human face area is illustrated as the specific image area, it is an example and is not limited thereto, and can be arbitrarily changed as appropriate, for example, a main area such as an object, a plant, or an animal. Also good.
As described above, the face identifying unit 3i includes identifying means for identifying a specific image region in the image based on the similarity calculated by at least one of the first calculation control unit 3d and the second calculation control unit 3h. It is composed.

  The memory 4 is constituted by, for example, a DRAM (Dynamic Random Access Memory) or the like, and temporarily stores data processed by the image processing unit 2, the face detection unit 3, the central control unit 8, and the like. Specifically, the memory 4 includes an image memory 4a and a storage memory 4b.

The image memory 4 a temporarily stores image data of an image to be displayed read from the image recording unit 6.
The storage memory 4b temporarily stores image data of an image area corresponding to the face area identified by the face identifying section 3i in the image to be displayed read from the image recording section 6.

The identification information recording unit 5 is constituted by, for example, a nonvolatile memory (flash memory). In addition, the identification information recording unit 5 records identification information for identifying a specific image area as a first recording unit. That is, the identification information recording unit 5 applies the specific image region to the image input to the face detection unit 3, specifically, the first region A1 and the second region A2 generated by the image region generation unit 3a. ID information for identifying a face is recorded.
Here, the identification information is, for example, whether or not the first area A1 and the second area A2 are input to the sub-classifier of the predetermined face detection unit 3 such as a neural network, Adaboost, or support vector machine. It is various information for determining.

The image recording unit 6 is configured by, for example, a nonvolatile memory (flash memory). The image recording unit 6 encodes and records image data of an image displayed on the display unit 1 by a predetermined encoding method as a second recording unit.
Note that the image data may be image data of a recording image captured at a predetermined size (for example, 10 million pixels), or reduced to a predetermined size (for example, VGA or QVGA size) for display. It may be image data of a display image.

  The operation input unit 7 has an operation unit constituted by, for example, data input keys for inputting numerical values, characters, etc., up / down / left / right movement keys for performing data selection, feeding operations, etc., various function keys, and the like. Then, a predetermined operation signal is output to the central control unit 8 according to the operation of these operation units.

  The central control unit 8 controls each unit of the image display device 100. Specifically, although not shown, the central control unit 8 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), and a ROM (Read Only Memory), and various processing programs for the image display device 100 ( Various control operations are performed according to (not shown).

Next, image identification processing by the image display apparatus 100 will be described with reference to FIGS.
FIG. 2 is a flowchart illustrating an example of an operation related to the image identification process. FIG. 3 is a flowchart illustrating an example of an operation related to the face detection process.
In the following description, an image P including a face area as a specific image area is used.

  As shown in FIG. 2, first, the central control unit 8 reads out image data of a desired image P to be displayed from the image recording unit 6 and temporarily stores it in the image memory 4a (step S1). Next, the image processing unit 2 acquires image data from the image memory 4a, decodes the image data in accordance with a corresponding predetermined encoding method (for example, JPEG format), and then has a predetermined size (for example, VGA or QVGA). The brightness image data L of the reduced image data is generated (step S2).

Subsequently, the face detection unit 3 performs face detection processing from the luminance image data L (step S3; see FIG. 3).
Here, the face detection process will be described in detail with reference to FIG.

As shown in FIG. 3, first, the face detection unit 3 uses a predetermined size (for example, 24 × 24) as a face identification target region by the evaluation value calculation unit 3 c from the luminance image data L generated by the image processing unit 2. It is determined whether or not the first area A1 of (pixel) can be generated (step S31). Specifically, the face detection unit 3 determines that the x-coordinate and y-coordinate of the pixel at the lower right corner of the sub-window A for setting the first area A1 in the luminance image data L are the lower right corner of the luminance image data L. It is determined whether or not the first region A1 can be generated depending on whether or not the pixel becomes larger than the x coordinate or the y coordinate.
Here, if it is determined that the first region A1 can be generated (step S31; YES), the image region generation unit 3a generates the first region A1 from the luminance image data L (step S32). Specifically, the image area generation unit 3a sets the reference blocks at the upper left corners of the first area A1 and the second area A2 to be arranged in a checkered pattern, and then the first in the luminance image data L. A first area A1 is generated by setting a subwindow A of a predetermined size (for example, 24 × 24 pixels) at a predetermined position with reference to each reference block (for example, the pixel at the upper left corner) related to the area A1 ( FIG. 5 and FIG. 6 (a) etc.). The face detection unit 3 temporarily stores the image data of the first area A1 generated by the image area generation unit 3a in the temporary storage memory 3b.

Next, the first calculation control unit 3d causes the evaluation value calculation unit 3c to calculate a similarity evaluation value related to the similarity between the first area A1 and the identification information recorded in the identification information recording unit 5 (Step S1). S33). Specifically, the evaluation value calculation unit 3c calculates a similar evaluation value according to the identification result of the sub classifier for the first area A1 using, for example, adaboost output calculation (see FIG. 7).
Subsequently, the face detection unit 3 temporarily stores the similarity evaluation value related to the first area A1 calculated by the evaluation value calculation unit 3c in the temporary storage memory 3b (step S34).

Thereafter, the face detection unit 3 can set the next first area A1 as a processing target (step S35), shift the process to step S31, and generate the first area A1 having a predetermined size from the luminance image data L. It is determined whether or not there is (step S31).
By repeatedly executing the above process until it is determined in step S31 that the first area A1 cannot be generated, the face detection unit 3 includes a plurality of first areas A1,. 6 (b), FIG. 6 (c), etc.) are sequentially set, and the similarity evaluation value is sequentially calculated for each first region A1.

When it is determined in step S31 that the first area A1 cannot be generated (step S31; NO), the face detection unit 3 uses the evaluation value calculation unit 3c to identify the face identification target area from the luminance image data L. It is determined whether or not the second area A2 having a predetermined size (for example, 24 × 24 pixels) can be generated (step S36). Note that the method of determining whether or not the second region A2 can be generated is substantially the same as the method of determining whether or not the first region A1 can be generated in step S31, and detailed description thereof is omitted.
If it is determined that the second area A2 can be generated (step S36; YES), the image area generation unit 3a generates the second area A2 from the luminance image data L (step S37). Specifically, the image area generation unit 3a uses the reference block at the upper left corner of the second area A2 arranged in a checkered pattern when the first area A1 is generated in the luminance image data L as a reference. The second area A2 is generated by setting the sub-window A having a size (for example, 24 × 24 pixels) at a predetermined position (see FIG. 5 and the like). The face detection unit 3 temporarily stores the image data of the second region A2 generated by the image region generation unit 3a in the temporary storage memory 3b.

Next, the determination unit 3e evaluates the similarity of the first area A1 including the reference block adjacent to the first area A1 in the vicinity of the second area A2 that is the identification target area, that is, the reference block of the second area A2. When the value (see FIG. 8) is added and the evaluation value for determination of the second area A2 (for example, the similarity evaluation value of the first area A1 is “2”, “2”, “7”, “7”) The evaluation value for determination of the second region A2 is calculated (step S38). Then, the determination unit 3e determines whether or not the calculated evaluation value for determination is greater than or equal to the evaluation threshold value (step S39).
Here, if it is determined that the evaluation value for determination is equal to or greater than the evaluation threshold (step S39; YES), the image region designating unit 3g evaluates the second region A2 in which the evaluation value for determination is equal to or greater than the evaluation threshold. A value is designated as an image area to be calculated (step S40).

Subsequently, the second calculation control unit 3h determines a similarity evaluation value related to the similarity between the second area A2 designated as the image area in step S40 and the identification information recorded in the identification information recording unit 5. Calculation is performed by the evaluation value calculation unit 3c (step S41). Specifically, the evaluation value calculation unit 3c calculates a similar evaluation value according to the identification result of the sub classifier for the second region A2, for example, using adaboost output calculation (see FIG. 7).
Subsequently, the face detection unit 3 temporarily stores the similarity evaluation value related to the second area A2 calculated by the evaluation value calculation unit 3c in the temporary storage memory 3b (step S42).

After that, the face detection unit 3 can set the next second area A2 as a processing target (step S43), shift the process to step S36, and generate a second area A2 of a predetermined size from the luminance image data L. It is determined whether or not there is (step S36).
Further, when it is determined in step S39 that the evaluation value for determination of the second area A2 serving as the identification target area is not equal to or higher than the evaluation threshold (step S39; NO), the face detection unit 3 performs the process in step S43. Then, the next second area A2 is set as a processing target (step S43), and the subsequent processing is executed. That is, for example, the second area A2 in the vicinity of the first area A1 is set in step S39 as in the case where the first area A1 is set in an image area having a small luminance change such as a blue sky or a plain background portion. Thus, it is determined that the evaluation value for determination of the second area A2 is not equal to or higher than the evaluation threshold (step S39; NO), and the subsequent processing is not performed.
By repeatedly executing the above processing until it is determined in step S36 that the second region A2 cannot be generated (step S36; NO), the face detection unit 3 includes a plurality of second images in the luminance image data L. The areas A2,... Are sequentially set, and it is sequentially determined whether or not the similarity evaluation value is calculated for each second area A2, and the similarity evaluation value is determined for the second area A2 that is determined to calculate the similarity evaluation value. Are calculated sequentially.

  When it is determined in step S36 that the second area A2 cannot be generated (step S36; NO), the face identifying unit 3i determines the similarity evaluation value of the first area A1 and the similarity evaluation value of the second area A2. Based on the above, a process of identifying a face area in the luminance image data L (step S44; see FIG. 9A) is performed, and the face detection process is terminated.

As shown in FIG. 2, the reduced image generation unit 2 a then sets the horizontal and vertical pixels of the image data based on the luminance image data L of the image P related to the display target to a predetermined ratio (for example, 0. It is determined whether reduced image data R (first reduced image data R1) reduced by (9 times) can be generated (step S4). Specifically, the reduced image generation unit 2a determines, for example, the number of pixels in the horizontal direction or the number of pixels in the vertical direction of the sub window A (for example, 24 × 24 pixels) related to the generation of the first region A1 and the second region A2. Also, whether or not the first reduced image data R1 can be generated is determined depending on whether the number of pixels in the same direction of the original image data (luminance image data L) is large.
Here, if it is determined that the reduced image data R (first reduced image data R1) can be generated (step S4; YES), the reduced image generation unit 2a stores the luminance image data L related to the display target image. After generating the first reduced image data R1 obtained by reducing the horizontal and vertical pixels at a predetermined ratio (for example, 0.9 times) (step S5), the process proceeds to step S3, and the face detection unit 3 Face detection processing is performed from the first reduced image data R1 (step S3; see FIG. 3).

The face detection process here is substantially the same as that described above with the luminance image data L as the image data to be processed except that the first reduced image data R1 is used as the image data to be processed. The detailed explanation is omitted.
In this face detection process, the face identifying unit 3i uses the first reduced image data R1 based on the similarity evaluation value of the first area A1 and the similarity evaluation value of the second area A2 generated from the first reduced image data R1. The face area is identified within the area (step S44; see FIG. 9A).

After the face detection process using the first reduced image data R1 as the processing target image data, the reduced image generation unit 2a performs the image data processing based on the first reduced image data R1 of the image P related to the display target. It is determined whether or not it is possible to generate reduced image data R (second reduced image data R2) obtained by reducing the horizontal and vertical pixels at a predetermined ratio (for example, 0.9 times) (step S4). Note that the method for determining whether or not the second reduced image data R2 can be generated is substantially the same as the method for determining whether or not the first region A1 can be generated, and detailed description thereof is omitted.
Here, if the reduced image generation unit 2a determines that the reduced image data R (second reduced image data R2) can be generated (step S4; YES), the reduced image generation unit 2a applies the image to be displayed. After generating the second reduced image data R2 obtained by reducing the horizontal and vertical pixels of the first reduced image data R1 at a predetermined ratio (for example, 0.9 times) (step S5), the process proceeds to step S3. The face detection unit 3 performs face detection processing from the second reduced image data R2 (step S3; see FIG. 3).

The face detection process here is substantially the same as that described above for the luminance image data L as the image data to be processed except that the second reduced image data R2 is used as the image data to be processed. The detailed explanation is omitted.
In this face detection process, the face identification section 3i uses the second reduced image data R2 based on the similarity evaluation value of the first area A1 and the similarity evaluation value of the second area A2 generated from the second reduced image data R2. The face area is identified within the area (step S44; see FIG. 9A).

The above processing is repeatedly executed until it is determined in step S4 that reduced image data R cannot be generated (step S4; NO), so that the face identification unit 3i can determine the similarity evaluation value of the first region A1 and the like. Based on the similarity evaluation value of the second area A2, a face area is identified in each reduced image data R (for example, the third to nth reduced image data R3,..., Rn).
As a result, as shown in FIG. 9A, the temporary face frame Wa is superimposed on each of the plurality of face areas identified in the luminance image data L and each reduced image data R. For example, a person who has a larger area of the entire image, such as a person on the front side in a group photo, or who is facing the front is smaller when generating reduced images sequentially. The face area can be identified even in the reduced image data R. This allows people who have a larger area of the entire image, or who are facing the front, to have a smaller area of the face, It becomes easy to identify the face area, and more temporary face frames Wa are superimposed.

Next, the face identifying unit 3i integrates at least one face area identified from the luminance image data L and the reduced image data R according to a predetermined arithmetic expression, and corresponds to one face area in the display target image. An image region is specified (step S6).
After that, the display control unit 1b displays the face area frame Wb superimposed on the image area (face area) identified by the face identification unit 3i in the display target image (step S7; FIG. 9B). )), The image region corresponding to the face region in the image recorded in the image recording unit 6 is subjected to a predetermined processing and displayed on the display panel 1a (step S8).
Specifically, for example, the display control unit 1b enlarges an image area corresponding to the face area identified by the face identification unit 3i at a predetermined magnification, and displays the image area on the display screen of the display panel 1a. The demosaic process is performed on the image area corresponding to the face area identified by the unit 3i and displayed on the display screen of the display panel 1a.
This completes the face identification process.

As described above, according to the image display device 100 of the present embodiment, the similarity between each of the first areas A1 generated from the image P and the identification information is calculated by the evaluation value calculation unit 3c. Based on the degree of similarity, an image area in which the degree of similarity with the identification information is calculated by the evaluation value calculation unit 3c is designated in each of the second areas A2. Then, the similarity between the designated image area and the identification information is calculated by the evaluation value calculation unit 3c, and based on the calculated similarity of at least one of the first area A1 and the second area A2. Since the specific image region (face region) is identified in the image P, it is not necessary to calculate the similarity with the identification information for all of the plurality of second regions A2,. The identification speed of the image area can be increased.
Specifically, the first area A1 and the second area A2 are generated so that the reference block composed of predetermined unit pixels of the first area A1 and the second area A2 is divided into direct sums. However, based on the similarity calculated for the first area A1, the image area for calculating the similarity with the identification information is specified among the plurality of second areas A2,. It is not necessary to calculate the similarity to the identification information for all of the areas A2,.

Further, based on the similarity calculated for the first area A1, it is determined whether or not the similarity with the identification information is calculated for each of the second areas A2, so that a plurality of second areas A2,. Among these, it is possible to designate an image region for calculating the similarity with the identification information. Specifically, an evaluation threshold value is set as a criterion for determining whether to calculate the similarity with the identification information for each of the second areas A2, and the similarity calculated for the first area A1 is When it is determined that the threshold value is equal to or greater than the evaluation threshold, it is determined that the similarity between the second area A2 and the identification information is calculated. Since it is determined that the similarity with the information for use is not calculated, it is possible to appropriately specify the second area A2 for calculating the degree of similarity with the information for identification.
In particular, by appropriately changing the value of the evaluation threshold as appropriate, the ratio designated as the image area for calculating the similarity to the identification information in all the second areas A2 can be changed. As described above, the identification accuracy of the specific image region in the image P can be changed as appropriate.

  Further, the first area A1 and the second area A2 are respectively generated from each of the reduced image data R generated by sequentially reducing the image P recorded in the image recording unit 6 at a predetermined ratio. Based on the similarity of at least one of the first area A1 and the second area A2 generated from the image data R, each specific image area can be identified, and at least one specific image area is predetermined. It is possible to specify an image area corresponding to one specific image area in the image P to be displayed by integrating according to the above equation. Therefore, the number of specific image areas identified can be increased or decreased by increasing or decreasing the number of reduced image data R generated, and the identification accuracy of specific image areas in the image P can be changed.

  Further, the image P recorded in the image recording unit 6 is displayed on the display unit 1, and more specifically, a predetermined processing process is performed on an image region corresponding to the specific image region identified in the image P. Since the specific image area cannot be identified, the image P to be displayed is simply reproduced and displayed. When the specific image area can be identified, the image corresponding to the specific image area is displayed. It is possible to reproduce and display an image obtained by performing predetermined processing on the area, and to realize a slideshow reproduction rich in expression.

The present invention is not limited to the above-described embodiment, and various improvements and design changes may be made without departing from the spirit of the present invention.
For example, the configuration of the image display apparatus 100 is merely an example of the configuration described in the above embodiment, and is not limited thereto. In addition, the image display device 100 is exemplified as the image identification device, but the image display device 100 is not limited thereto. For example, as illustrated in FIG. Any configuration can be used as long as it can execute image processing.
Hereinafter, an imaging apparatus 200 as a modification of the image identification apparatus will be described with reference to FIG.

As illustrated in FIG. 10, the imaging apparatus 200 includes a display unit 1, an image processing unit 2, a face detection unit 3, a memory 4, an identification information recording unit 5, an image recording unit 6, and an operation input unit. 7 and the central control unit 8, an imaging unit 9 is provided.
The display unit 1, the image processing unit 2, the face detection unit 3, the memory 4, the identification information recording unit 5, the image recording unit 6, the operation input unit 7, and the central control unit 8 are as follows. In other respects, the configuration is substantially the same as that of the image display device 100 of the above embodiment.

  The imaging unit 9 constitutes an imaging unit that captures an image. Specifically, although not shown, the imaging unit 9 includes a lens unit composed of a plurality of lenses such as a zoom lens and a focus lens, a diaphragm for adjusting the amount of light passing through the lens unit, and a CCD ( An electronic imaging unit, which is composed of an image sensor such as a charge coupled device (CMOS) or a complementary metal-oxide semiconductor (CMOS), converts an optical image that has passed through various lenses of the lens unit into a two-dimensional image signal, a timing generator, and a driver And an imaging control unit composed of the above. Then, the imaging control unit scans and drives the electronic imaging unit with a timing generator and a driver, converts an optical image formed by the lens unit at predetermined intervals into a two-dimensional image signal with the electronic imaging unit, and Image frames are read out from the imaging area of the electronic imaging unit for each screen and are output to the image processing unit 2.

  The image processing unit 2 appropriately adjusts the gain for each of the RGB color components with respect to the analog value signal of the image data transferred from the electronic imaging unit, and then performs sample holding by a sample hold circuit (not shown) to perform A / D. After conversion into digital data by a converter (not shown) and color process processing including pixel interpolation processing and γ correction processing by a color process circuit (not shown), the luminance signal Y and color difference signals Cb, Cr of digital values (YUV image data) is generated.

  The image recording unit 6 encodes and records the YUV image data generated by the image processing unit 2 using a predetermined encoding method (for example, JPEG format).

  Therefore, according to the imaging apparatus 200 described above, the specific image area (face area) can be identified from the image recorded in the image recording unit 6, that is, the image captured by the imaging unit 9. As a result, the image corresponding to the identified specific image region can be displayed after being subjected to predetermined processing such as enlargement processing or demosaic processing.

  In the imaging apparatus 200 described above, the specific image area is identified from the image recorded in the image recording unit 6. However, the specific image area is identified from the live view image generated by imaging the subject by the imaging unit 9. The image area may be identified, and this can improve the detection speed of the specific image area (face area) when the subject is imaged. Further, when the automatic focusing process (AF) is performed using the detected specific image area, the AF speed can be improved.

  In the above embodiment, the sizes of the first region A1 and the second region A2 are made equal. However, the present invention is not limited to this, and the sizes of the first region A1 and the second region A2 are different. May be allowed. For example, the size of the second area A2 may be relatively larger than the first area A1.

In addition, in the above embodiment, the functions as the generation unit, the first calculation control unit, the designation unit, the second calculation control unit, and the identification unit are detected under the control of the central control unit 8. Although the configuration realized by driving the unit 3 is not limited thereto, the configuration may be realized by executing a predetermined program or the like by the CPU of the central control unit 8.
That is, a program including a generation process routine, a first calculation control process routine, a designation process routine, a second calculation control process routine, and an identification process routine is stored in a program memory (not shown) that stores the program. Then, the CPU of the central control unit 8 may function as a generation unit that generates a plurality of first areas A1 and second areas A2 from the image recorded in the second recording unit by the generation processing routine. . Also, a first calculation control unit that causes the CPU of the central control unit 8 to calculate the similarity between each of the first areas A1 generated by the generation unit and the identification information by the first calculation control processing routine. You may make it function as. In addition, the CPU of the central control unit 8 by the designation processing routine makes the identification information out of each of the second areas A2 generated by the generation unit based on the similarity calculated by the first calculation control unit. You may make it function as a designation | designated means to designate the image area | region which calculates a similarity by a calculation means. Further, the CPU of the central control unit 8 is caused to function as a second calculation control unit that causes the calculation unit to calculate the similarity between the image area specified by the specifying unit and the identification information by the second calculation control processing routine. Anyway. Further, the CPU of the central control unit 8 identifies a specific image area in the image based on the similarity calculated by at least one of the first calculation control means and the second calculation control means by the identification processing routine. You may make it function as an identification means to do.

  Similarly, the determination unit, the setting unit, the image reduction unit, and the display control unit may be realized by executing a predetermined program or the like by the CPU of the central control unit 8.

  Furthermore, as a computer-readable medium storing a program for executing each of the above processes, a non-volatile memory such as a flash memory or a portable recording medium such as a CD-ROM is applied in addition to a ROM or a hard disk. Is also possible. A carrier wave is also used as a medium for providing program data via a predetermined communication line.

DESCRIPTION OF SYMBOLS 100 Image display apparatus 1 Display part 1a Display panel 1b Display control part 2 Image processing part 2a Reduced image generation part 3 Face detection part 3a Image area generation part 3c Evaluation value calculation part 3d First calculation control part 3e Determination part 3f Evaluation threshold value setting Unit 3g image region designation unit 3h second calculation control unit 3i face identification unit 5 identification information recording unit 6 image recording unit 8 central control unit 9 imaging unit 200 imaging device

Claims (9)

First recording means for recording identification information for identifying a specific image region;
A second recording means for recording an image;
Generating means for generating a plurality of first areas and second areas from the images recorded in the second recording means;
Calculating means for calculating the similarity between the image recorded in the second recording means and the identification information recorded in the first recording means;
First calculation control means for causing the calculation means to calculate the similarity between each of the first areas generated by the generation means and the identification information;
Based on the similarity calculated by the first calculation control means, an area in which the calculation means calculates the similarity to the identification information among each of the second areas generated by the generation means. A designation means to designate,
Second calculation control means for causing the calculation means to calculate the similarity between the area specified by the specification means and the identification information;
Identification means for identifying the specific image region in the image based on the similarity calculated by at least one of the first calculation control means and the second calculation control means;
Based on the similarity calculated by the first calculation control means and the similarity of the plurality of the first areas around the second area, the identification information for each of the second areas Determination means for determining whether or not the calculation means calculates the similarity;
With
The designation means is:
An image identification apparatus, wherein , based on a determination result by the determination unit, an area in which the similarity with the identification information is calculated by the calculation unit is specified out of each of the second areas . A setting unit that sets a threshold value that is a determination criterion by the determination unit as to whether or not the calculation unit calculates the similarity to the identification information for each of the second regions;
The determination means further includes:
When it is determined that the similarity calculated by the first calculation control unit and the similarities of the plurality of first regions around the second region are greater than or equal to a threshold set by the setting unit In addition, it is determined that the similarity with the identification information is calculated by the calculation unit, and the similarity is calculated by the first calculation control unit , and a plurality of the first regions around the second region are calculated . claim 1, similarity 1 region when it is determined to be less than the set threshold value by said setting means, characterized in that it determines not to calculate the degree of similarity between the identifying information by said calculation means The image identification device described in 1. Image reduction means for sequentially reducing the images recorded in the second recording means at a predetermined ratio and sequentially generating reduced images;
It said generating means, image identification apparatus according to claim 1 or 2, characterized in that respectively generate the first and second regions from each of the reduced images are sequentially generated by the image reduction unit. The generation unit generates the first area and the second area, respectively, so that a reference block composed of predetermined unit pixels in each of the first area and the second area is divided into direct sums. Item 4. The image identification device according to any one of Items 1 to 3 . Image identification apparatus according to any one of claims 1-4, characterized by further comprising display means for displaying the image recorded in the second recording means. 6. The display control unit according to claim 5 , further comprising a display control unit configured to perform a predetermined processing on an image corresponding to the specific image area identified by the identification unit in the image and display the image on the display unit. The image identification device described. It further comprises imaging means for taking an image,
It said second recording means, image identification apparatus according to any one of claim 1 to 5, characterized in that to record the image captured by the imaging means. First recording means for recording identification information for identifying a specific image area, second recording means for recording an image, an image recorded in the second recording means, and the first recording An image identification method using an image identification device comprising: a calculation means for calculating a similarity with the identification information recorded in the means,
A generation step of generating a plurality of first areas and a plurality of second areas from the image recorded in the second recording means;
A first calculation control step for causing the calculation means to calculate the similarity between each of the generated first areas and the identification information;
Based on the similarity calculated in the first calculation control step, a designation step for designating an area in which the similarity with the identification information is calculated by the calculation means among each of the second areas;
A second calculation control step for causing the calculation means to calculate the similarity between the area specified in the specifying step and the identification information;
Identifying the specific image area in the image based on the calculated similarity of at least one of the first area and the second area;
The similarity calculated by the first calculation control step, and based on the similarity of the plurality of the first areas around the second area, the identification information for each of the second areas A determination step of determining whether or not the similarity is calculated by the calculation unit;
Including
The specifying step includes:
An image identification method, wherein , based on the determination result of the determination step, an area for calculating the similarity with the identification information among each of the second areas is specified. First recording means for recording identification information for identifying a specific image area, second recording means for recording an image, an image recorded in the second recording means, and the first recording A computer of an image identification apparatus comprising: a calculation unit that calculates a similarity to the identification information recorded in the unit;
Generating means for generating a plurality of first areas and second areas from images recorded in the second recording means;
First calculation control means for causing the calculation means to calculate the similarity between each of the first regions generated by the generation means and the identification information;
Based on the similarity calculated by the first calculation control means, an area in which the calculation means calculates the similarity to the identification information among each of the second areas generated by the generation means. Designation means to designate,
Second calculation control means for causing the calculation means to calculate the similarity between the area specified by the specification means and the identification information;
Identification means for identifying the specific image region in the image based on the similarity calculated by at least one of the first calculation control means and the second calculation control means;
Based on the similarity calculated by the first calculation control means and the similarity of the plurality of the first areas around the second area, the identification information for each of the second areas Determination means for determining whether or not the calculation means calculates the similarity;
Function as
The designation means is:
A program for designating a region in which the similarity with the identification information is calculated by the calculating unit among the second regions based on the determination result by the determining unit .

Images